Author + information
- Received January 25, 2013
- Revision received May 28, 2013
- Accepted June 3, 2013
- Published online August 27, 2013.
- Stéphane Lafitte, MD, PhD∗,
- Patricia Reant, MD, PhD∗,
- Cecile Touche, MD∗,
- Xavier Pillois, PhD∗,
- Marina Dijos, MD∗,
- Florence Arsac, MD∗,
- Jerome Peyrou, MD∗,
- Michel Montaudon, MD, PhD∗,
- Philippe Ritter, MD∗,
- Raymond Roudaut, MD∗ and
- Anthony DeMaria, MD, PhD†∗ ()
- ∗University of Bordeaux, Centre Hospitalier Universitaire de Bordeaux, Bordeaux, France
- †Sulpizio Cardiovascular Center, University of California at San Diego, San Diego, California
- ↵∗Reprint requests and correspondence:
Dr. Anthony DeMaria, Division of Cardiology, University of California, San Diego, 3600 Campus Point Drive, San Diego, California 92103.
Objectives The purpose of this study was to analyze left ventricular obstruction in patients with hypertrophic cardiomyopathy (HCM) during exercise echocardiography.
Background Despite the association of symptoms with left ventricular outflow tract obstruction in HCM, there exist paradoxical situations in which significant intraventricular gradients (>50 mm Hg) at rest occur in conjunction with excellent exercise tolerance.
Methods To examine this phenomenon, we performed exercise echocardiography and analyzed the clinical status of 107 HCM patients with and without resting obstruction.
Results At rest, 69 patients had no obstruction and 38 exhibited an intraventricular gradient, 9 of whom exhibited a decrease in gradient of at least 30 mm Hg (99.4 ± 35.5 mm Hg to 30.2 ± 14.3 mm Hg, p < 0.001) during exercise (paradoxical response to exercise [PRE]). The PRE patients presented with a significantly lower New York Heart Association clinical class and higher left ventricular volumes and arterial pressure both at rest and during exercise than HCM patients in whom the gradient increased or did not change during stress echocardiography. Finally, PRE patients exhibited a trend toward a reduced rate of cardiac events.
Conclusions Our study reports a subgroup of HCM patients, designated PRE based on a decreased intraventricular gradient during exercise. The reduced exertional obstruction may account for the better functional class and trend toward fewer clinical events in PRE patients.
Hypertrophic cardiomyopathy (HCM) is the most common heritable cardiovascular disorder, affecting 0.2% to 0.5% of the general population, and is a leading cause of sudden cardiac death in young athletes (1). HCM is characterized by an increased thickness of heart muscle, typically consisting of asymmetric septal hypertrophy, and systolic anterior motion (SAM) of the mitral valve. Resting left ventricular outflow tract obstruction (LVOTO) due to SAM is observed in 25% to 30% of HCM patients and, when severe, may cause dyspnea, chest pain, syncope, and a predisposition to developing atrial arrhythmias (2).
Insights into the pathophysiology of LVOTO have recently been provided by exercise echocardiography, which can quantify the gradient during or after exercise (3–6). Although the majority of HCM patients with resting obstruction have increasing symptoms with exertion (3), presumably in proportion to worsening LV obstruction, some surprisingly do not. Despite the presence of resting gradients exceeding 60 mm Hg in these unusual HCM patients, they are unexpectedly often able to perform exercise treadmill testing to levels of 150 W and greater.
On the basis of these clinical observations, we hypothesized that some HCM patients may have a paradoxical response to exercise (PRE), consisting of a decrease of LVOTO during exercise that results in a preservation of effort tolerance. Therefore, we conducted a prospective exercise echocardiography study of HCM patients for whom this procedure was scheduled for clinical indications. We divided HCM patients into 2 subgroups, namely, patients whose gradient either increased or did not significantly change and patients whose gradient decreased, and analyzed the exercise tolerance, clinical status, and echocardiographic measures of LV structure and function for the individual subgroups.
From May 2009 to December 2011, 120 patients with HCM were referred to evaluate functional capacity using exercise echocardiography. The medical unit from which the patients were drawn employs a specialized team, including cardiac and genetic physicians, to manage >200 HCM patients. This team keeps detailed clinical records for patient management. The HCM was diagnosed by conventional criteria (7,8). The study inclusion criteria were as follows: 1) referral for exercise echocardiography; 2) previous formal diagnosis of HCM based on both morphological hypertrophy and familial history; 3) sinus rhythm; and 4) ability to perform bicycle exercise testing. Exclusion criteria were: 1) poor ultrasonography window quality; and 2) recent history of syncope, chest pain, or severe arrhythmia during exercise. Information regarding the study and data collection was provided to all patients, and the protocol was approved by the institutional review board.
Resting 2-dimensional (2D) echocardiography was performed according to American Society of Echocardiography guidelines (9,10), with ultrasound recordings obtained on a Vivid 9-dimension ultrasound system (General Electric Medical System, Horten, Norway) by an experienced (level 3) operator (11). Recordings in standardized views were acquired in 2D, pulsed, continuous, and color Doppler modalities and stored for subsequent analysis. Particular attention was paid to the LVOT area to identify and analyze systolic anterior motion of the mitral valve in both the parasternal long-axis and apical 3- and 5-chamber views. After scanning the LVOT with continuous wave Doppler, the maximal outflow velocity was found and was measured. Outflow gradients were automatically calculated from the flow velocities using the modified Bernoulli equation (12). Heart rate was calculated at the time of gradient measurement based on the average of 3 R-R intervals.
The indications for exercise echocardiography were as follows: evidence of latent LVOTO in 64% of patients, clinical discordance between symptoms and signs in 10%, evaluation of medical treatment responses in 20%, and evaluation of mitral regurgitation during effort in 6%. Functional status and blood pressure adaptation during exercise were assessed in all patients.
Exercise echocardiography was performed without stopping medications, according to the European Association of Echocardiography guidelines (13), as bicycle exertion in a semisupine position (50°) with a slight left lateral tilt so as to enable simultaneous transthoracic echocardiography during exercise. Starting at 25 W, the work load was increased by 25 W every 2 min to the maximum tolerated effort. At each stage from rest to recovery, electrocardiographic and conventional echocardiography recordings in 2D views and continuous and color Doppler modalities were acquired and stored for of-fline analysis. Blood pressure was measured using a cuff sphygmomanometer at rest, at 1-min intervals during exercise, and at 1-min intervals for 5 min during the recovery period after exercise (14).
An independent observer blinded to patient history analyzed all cases retrospectively, applying standard measurements according to European Association of Echocardiography/American Society of Echocardiography guidelines and using the internal quantitation package of the echocardiograph. Localization of hypertrophy was performed according to the classification of Maron et al. (15). An estimate of LV mass was obtained from the M-mode tracing. The LV and left atrial volumes were calculated from the apical 2- and 4-chamber views using Simpson's rule (9). Aortic and mitral velocities and velocity–time integral were obtained using pulsed Doppler, and pulmonary pressure was calculated from the measured tricuspid gradient as (4 [tricuspid velocity (in m/s)]2) plus an estimate of right atrial pressure from the inferior vena cava (16). Filling pressures were estimated from the ratio of mitral orifice and septal annular early diastolic (E) velocities as E/E′ ratio. Longitudinal LV deformation with contraction was obtained using the 2D speckle tracking method (17). The length of the anterior mitral leaflet was measured from the parasternal long-axis view during mid-diastole and with the leaflet maximally extended as the distance from the junction between the anterior leaflet and the posterior aortic wall to the tip of the leaflet (18). Distance from SAM to the septum as well as duration of mitral/septum contact when adequately defined were measured at rest and during exercise using anatomic M-mode (19).
Because latent obstruction can occur at multiple LVOT points, outflow velocities were measured using continuous wave Doppler during exercise with the same direction and angle as recorded at rest, and the highest outflow gradient measured during the entire test was recorded. Specific attention was paid so as not to confuse with mitral regurgitation flow when present.
Medical history, other investigations, and cardiac events
Each patient's medical history was extracted from the institutional database, including date of HCM diagnosis, family history, symptoms and complications, pacemaker implantation or heart surgery, and treatment at the time of the study. Results of standardized cardiac magnetic resonance imaging, 24-h Holter electrocardiography, and biomarker parameters, such as B-type natriuretic peptide or N-terminal pro–B-type natriuretic peptide, were also extracted when available, provided that the tests were performed within a year before exercise echocardiography.
All data were expressed as mean ± SD, and comparison was performed between subgroups using Student's unpaired t test for unpaired data and Fisher's test as appropriate. Independent variables of PRE were sought using multiple linear regression analysis.
Specific data from this study are provided in Tables 1 to 3. Of the 120 consecutive patients with known HCM referred for exercise echocardiography, 13 were excluded from analysis (7 for poor ultrasound window quality, 4 for atrial fibrillation, and 2 New York Heart Association [NYHA] functional class IV), resulting in 107 HCM patients in the study group. Patient characteristics are in Table 1. The mean age of the study group at inclusion was 51.6 ± 15.4 years, and their mean age at HCM diagnosis was 43.4 ± 18.5 years. Mean NYHA functional class was 1.7 ± 0.6. Among these patients, 72.9% were receiving beta-blockers, and 21.5% had undergone implantable cardioverter-defibrillator insertion in accordance with guideline recommendations (7): 9% alcohol septal ablation and 3.7% surgical septal reduction. Patient characteristics of the individual subgroups are presented in Table 1.
Resting echocardiographic measurements
The study population's mean maximal LV septal wall thickness was 21.1 ± 4.9 mm (Table 2). Using criteria based on Doppler measurements (i.e., outflow tract gradient exceeding 30 mm Hg), 38 (36%) patients had LVOT obstruction (OB) at rest and 69 (64%) patients were nonobstructed (NOB). The group with resting OB displayed a significantly higher maximal LV wall thickness than the NOB group (22.7 ± 4.8 vs. 20.3 ± 4.8, p = 0.016). We also observed a greater anterior mitral leaflet length associated with an increased mitral valve regurgitation grade in the OB group compared to the NOB group, as well as a larger left atrial volume.
All exercise echocardiographic studies were performed without complications; the primary reasons for exercise discontinuation were lower limb fatigue (73%) and dyspnea (21%). One patient had atrial tachycardia during exercise testing. Mean exercise performance level was 121 ± 41 W, with a significantly higher performance achieved in the group without resting obstruction as compared to the obstructed group. Mean systolic blood pressures did not differ at rest or during exercise between the NOB and OB groups. Although no significant difference in resting heart rate was noted between the 2 groups, OB patients had a lower exercise heart rate than NOB patients.
In the OB group (n = 38), the maximal gradient increase with exercise from 68.8 ± 33.2 mm Hg to 78 ± 58 mm Hg was nonsignificant because of major differences in the response of individual patients, as already described (20). Based on the response to exercise, patients in the OB cohort were divided into those who exhibited either a significant increase (>30 mm Hg [n = 15, from 72 ± 27 mm Hg to 138 ± 45 mm Hg]) or no to minimal change (<30 mm Hg [n = 14, from 44 ± 16 mm Hg to 43 ± 19 mm Hg]) in LVOT gradient (OBI group), and those in whom there was a significant decrease (>30 mm Hg [n = 9, from 92.4 ± 59.4 mm Hg to 30.2 ± 14.3 mm Hg]) in gradient (PRE group) (Fig. 1, Table 3).
Relation to symptoms/exercise tolerance
The OB group overall displayed a significantly higher mean NYHA class than the NOB group (1.9 ± 0.7 vs. 1.6 ± 0.6, p= 0.008), in addition to a significantly lower mean exercise tolerance (106 ± 29 W vs. 129 ± 45 W, p < 0.05). The majority of OB patients (73.7%) were symptomatic (NYHA class >1). Within the OB group, a significantly better mean NYHA class was observed in the PRE subgroup compared to the OBI subgroup (1.4 ± 0.5 vs. 2.1 ± 0.7, p = 0.009), as was a numerically but not significantly higher maximal exercise load (115 ± 30 W vs. 103 ± 28 W, p = NS). Systolic blood pressures were significantly higher at rest and during exercise in the PRE than in the OBI patients. The average increases in LVOT gradient from rest to exercise for patients with NYHA class I, II, and III were 4.3 ± 40 mm Hg, 20 ± 46 mm Hg, and 29 ± 24 mm Hg, respectively (p < 0.05 between NYHA class I and II). However, no correlation was found between NYHA class and the magnitude of increase of gradient with exercise.
Cardiac structure and function
At rest, the PRE group exhibited significantly higher end-diastolic and end-systolic LV volumes (43.8 ± 7.5 ml/m2 vs. 36.3 ± 8.6 ml/m2, and 14.4 ± 3.8 ml/m2 vs. 10.2 ± 3.6 ml/m2, respectively; both, p < 0.05), and a numerically but not significantly lower ejection fraction compared to the OBI patients (Table 3). The LV mass and longitudinal contraction were similar for these cohorts. Regarding diastolic function, only Doppler mitral inflow E/A velocities ratio statistically differed between the 2 groups. Finally, LVOTO was higher in the PRE group (99.4 ± 35.5 mm Hg vs. 59.2 ± 26.4 mm Hg, p= 0.003) despite a shorter anterior mitral valve (25.6 ± 1.2 mm vs. 29.7 ± 3.4 mm, p < 0.001). At rest, mitral to septum distance and duration of mitral contact to septum were similar in the 2 groups (respectively, 1.1 ± 2.7 mm vs. 1.3 ± 3.7 mm, p = NS; and 94.1 ± 64.5 ms vs. 81.1 ± 56.7 ms, p = NS). Interestingly, at-rest systolic blood pressure was higher in the PRE group (140 ± 18.8 mm Hg vs. 121.1 ± 14.5 mm Hg, p = 0.016), whereas heart rates were numerically but not significantly less than in the OBI group.
During exercise, mean gradients at peak exercise were 30.2 ± 14.3 mm Hg and 92.4 ± 59.4 mm Hg for the PRE and OBI patients, respectively (p = 0.004), and 140 ± 70 mm Hg and 114 ± 69 mm Hg in the first 30 s during recovery, respectively (p = 0.32). From rest to exercise, mitral to septum distance significantly increased in the PRE group (from 1.1 ± 2.7 mm to 7.8 ± 3.2 mm, p < 0.01) but not in the OBI group (p < 0.001 between groups). Duration of mitral contact to septum was similar in the OBI group (from 81.1 ± 56.7 ms to 77.6 ± 61.7 ms, p = NS) but was no longer present in the PRE group (p < 0.001 between groups).
Patients reached 68.7% and 70.6% (p = NS) of maximal heart rate in the PRE group and the OBI group, respectively, with a trend toward higher exercise workload in the PRE group (115 ± 30 W vs. 103 ± 28 W; p = NS). Systolic blood pressure during exercise was significantly higher in the PRE group, with a significantly higher systolic blood pressure increase during exercise (38 ± 10% vs. 25 ± 17%, p < 0.05) in the PRE group. At peak exercise, PRE patients exhibited higher end-diastolic LV volume and end-systolic LV volume (44.1 ± 8.5 ml/m2 vs. 33.2 ± 8.6 ml/m2, and 12.2 ± 2.9 ml/m2 vs. 9.7 ± 2.8 ml/m2, respectively; both, p < 0.05). Importantly, in the PRE group, exercise resulted in a significant reduction in end-systolic LV volume, with maintained end-diastolic LV volume. In the OBI group, these parameters remained unchanged during exercise. Thus, there was a trend toward higher ejection fraction levels in the PRE group (72.4 ± 4.0% vs. 69.8 ± 8.3%), but the between-group differences did not reach statistical significance, showing a correct adaptation of LV function.
Relation to Patients' History and Cardiac Events (N = 107)
Regarding HCM history (follow-up duration: 10.4 ± 9.4 years), a positive diagnosis was established at a more advanced age in the PRE group compared to the OBI group (57.6 ± 10.6 years vs. 42.5 ± 17.5 years, p = 0.025), which was associated with late symptoms and a lower functional NYHA class (1.4 ± 0.5 vs. 2.1 ± 0.7, p = 0.009). Similarly, the death risk criteria sum (premature HCM-related sudden death of ≥1 relatives; history of unexplained syncope judged inconsistent with neurocardiogenic origin; multiple and/or prolonged runs of nonsustained ventricular tachycardia on serial 24-h ambulatory Holter electrocardiographic monitoring at heart rates ≥120 beats/min; hypotensive or attenuated blood pressure response to exercise; and massive left ventricular hypertrophy ) was numerically lower in the PRE group (0.5 ± 0.7 vs. 1.0 ± 1.1, p = NS), with 11% of PRE patients having undergone implantable cardioverter-defibrillator insertion versus 34% of other OB patients.
Heart Morphology from MRI
In total, 31 of 38 OB patients (81%) underwent MRI examination. In the PRE group, there was a trend toward higher end-diastolic LV volume as compared to OBI patients (129 ± 63 vs. 114 ± 32, p = 0.2), along with a significantly lower maximum wall thickness (19 ± 2 mm vs. 23 ± 5 mm, p < 0.05). Late enhancement was found in 50% of PRE patients versus 75% of OBI patients.
Since the first demonstration of a relationship between the degree of LV obstruction and prognosis in HCM patients, interest in understanding the pathophysiology of this phenomenon has grown (22). Provoked latent obstruction has been invoked as a major cause of exercise intolerance (6). We analyzed 107 HCM cases referred for stress echocardiography, and identified a previously unappreciated response of LV obstruction to exertion characterized by a progressive decrease in LVOT gradient from rest to peak exercise. This phenomenon was observed in 8% of our HCM patients overall, and 23% of those with resting obstruction >30 mm Hg. We have referred to this phenomenon as paradoxical response to exercise (PRE), highlighting the unexpected reduction in gradient with exertion in HCM patients. This pattern of response to exertion was associated with trends toward a greater exertional capacity and a superior NYHA functional class.
Exercise echocardiography in HCM and PRE
Numerous recent publications have already addressed the questions of prevalence, significance, origin, and consequences of LVOTO during exercise in HCM, comparing patients without obstruction at rest with those presenting significant gradients before exercise testing. Different types of “stressors” have been evaluated, including pharmacological stimulation, Valsalva maneuvers, and exercise tests (23). Most studies investigating dynamic obstruction focused on conventional treadmill exercise followed by echocardiographic recordings at recovery (3–5,24,25). This approach cannot be considered to be a pure evaluation of exercise dynamics, as dramatic pre-load variations are observed a few seconds after the end of effort, especially when using a bicycle test. When upright, at the end of lower limb muscle exercise, there is a large decrease in venous blood return to the heart, yielding: 1) decreased LV volume; 2) decreased wall stress; 3) continued sympathetic drive; and 4) a hyperkinetic state like that observed during dobutamine-induced stress. Indeed, dobutamine infusion has been shown to generate obstructions independently of HCM (26,27), as can the post-exercise test (20).
In our study, we used bicycle exercise, simulating day-to-day exertional provocation and allowing for real-time screening of hemodynamics throughout the evaluation. Consequently, our findings are not fully comparable to those of immediate post–upright exercise studies. This crucial difference may also explain why other researchers have not previously described PRE in their series. However, 3 recent publications reported exercise studies. Shah et al. (4) and Nistri et al. (28) reported latent obstruction in 60% and 40% of patients, respectively; they were not able to observe any PRE, as resting obstruction was an exclusion criterion. The most recent study, by Jensen et al. (23), evaluated patients treated with septal myocardial ablation. Except for 7 patients, none of the others exhibited a resting gradient. During exercise, the gradient in these 7 patients increased to an average of 66 mm Hg.
Mechanisms of LV obstruction in HCM and PRE
Although the Venturi effect was believed to be responsible for SAM (4,29,30), the most recent evidence for LV obstruction in HCM patients favors the flow drag mechanism causing the mitral valve to be pushed against the septum (7). The mechanism of obstruction is probably also related to other alterations produced by HCM. Experimental and observational data suggest that anterior displacement of the papillary muscles and submitral apparatus is necessary to create sufficient leaflet slack to allow for anterior motion of the mitral leaflet (31,32). Nevertheless, the major potential effects of ventricular loading and myocardial contractility must also be considered. These effects may be exerted both in early systole, for which flow, drag, and pushing force of flow are the dominant hydrodynamic forces for SAM, and at midsystole, for which the displacing force is more prominent (5). Hence, small variations in preload, afterload, or contractility, such as produced by exertion, may lead to large changes in gradient, usually explaining the amplification of obstruction from rest to exercise or from exercise to recovery. Surprisingly, in a small group of patients, we observed an unexpected and significant decrease in the obstruction from rest to exercise, described here as PRE. The LVOTO decrease was associated with reduction in SAM (Fig. 1).
Comparisons of OBI and PRE patient groups revealed that load conditions significantly differed between the 2 patient groups. Notably end-diastolic LV volume and systolic blood pressure were greater both at rest and during exercise in PRE. We also observed more pronounced changes in these parameters from rest to exercise in the PRE group. Differences in mitral-septal contact accompanied these findings. These observations are consistent with the concept that specific characteristics of ventricular loading both at rest and during exertion may be responsible for the decrease in LVOTO with exercise. It is possible that the larger LV volumes may predispose to a decrease in obstruction because there is less slack in the submitral apparatus. Similarly, the decrease in systolic volume may be related to the reduction of the LVOTO.
Safety and prognosis
Although this study was not designed to prospectively evaluate the risk for complications in our HCM patient group, we retrospectively examined patients' records from our local database. Surprisingly, we observed a lower rate of recorded abnormalities, adverse events, and medical interventions such as implantable cardioverter-defibrillator insertion or septal ablation in the PRE patients. However, statistically significant levels were not achieved for all criteria, probably because of the limited size of the PRE group. Nonetheless, our findings demonstrate that exercise echocardiography in HCM patients with resting obstruction may be useful in identifying a subgroup of patients with a lesser risk of complications. The data also suggest that the effects on exertional capacity of pharmacologic agents should be tested in individual HCM patients, as these agents may potentially exert detrimental effects, especially in the PRE subset. Finally, our study confirmed previous data indicating the safety of exercise echocardiography in HCM patients, even in the presence of significant obstruction at rest (33,34).
The sample size of our study was limited, reflecting the clinical prevalence of HCM. Our study was not designed to assess the prevalence of PRE in HCM patients, as we included only patients referred to our echocardiography department for exercise echocardiography. Therefore, the proportion of PRE patients observed in our study cannot be extrapolated to the whole HCM population. We did not use objective means such as oxygen consumption (VO2) to quantify exercise capacity, and the examiners were not blinded during the performance of the stress test. The patients were evaluated in a semisupine position, which is not similar to treadmill exercise testing. Theoretically, PRE might be more common with exercise treadmill echocardiography as the patient is placed supine at termination and venous return may be augmented even more than with upright bike exercise, which is consistent with a better NYHA class in PRE during their routine life. Finally, as this is a retrospective study, we cannot establish a causal relationship between the response of the LVOT gradient to exercise and symptomatic status.
Left ventricular outflow obstruction is a central feature of the pathophysiology of hypertrophic cardiomyopathy, and is believed to increase during exercise with an attendant production of symptoms. This report describes a possible subgroup of HCM patients in whom a high resting gradient was reduced during exertion. This response was associated with a trend toward increased exertional capacity and functional class, fewer complications, and a more physiological profile of LV structure and function. If these findings are confirmed in larger studies, they may have important implications for patient management.
Dr. Ritter has a relationship with Sorin CRM and Medtronic and receipt of research grants. All other authors have reported they have no relationships relevant to the contents of this paper to disclose. Martin S. Maron, MD, served as Guest Editor for this article. The first 2 authors contributed equally to this study.
- Abbreviations and Acronyms
- hypertrophic cardiomyopathy
- left ventricular
- left ventricular outflow tract
- left ventricular outflow tract obstruction
- nonobstructive hypertrophic cardiomyopathy at rest
- New York Heart Association
- obstructive hypertrophic cardiomyopathy at rest
- obstructive hypertrophic cardiomyopathy at rest with increased gradient or no change during exercise
- paradoxical response to exercise
- systolic anterior motion
- Received January 25, 2013.
- Revision received May 28, 2013.
- Accepted June 3, 2013.
- American College of Cardiology Foundation
- Maron B.J.
- Maron M.S.,
- Olivotto I.,
- Zenovich A.G.,
- et al.
- Shah J.S.,
- Esteban M.T.,
- Thaman R.,
- et al.
- Gersh B.J.,
- Maron B.J.,
- Bonow R.O.,
- et al.
- McKenna W.J.,
- Elliott P.M.
- Quinones M.A.,
- Douglas P.S.,
- Foster E.,
- et al.
- Yoganathan A.P.,
- Cape E.G.,
- Sung H.W.,
- Williams F.P.,
- Jimoh A.
- Sicari R.,
- Nihoyannopoulos P.,
- Evangelista A.,
- et al.
- Ciampi Q.,
- Betocchi S.,
- Lombardi R.,
- et al.
- Serri K.,
- Reant P.,
- Lafitte M.,
- et al.
- Klues H.G.,
- Proschan M.A.,
- Dollar A.L.,
- Spirito P.,
- Roberts W.C.,
- Maron B.J.
- Pollick C.,
- Morgan C.D.,
- Gilbert B.W.,
- Rakowski H.,
- Wigle E.D.
- Klues H.G.,
- Leuner C.,
- Kuhn H.
- Pellikka P.A.,
- Oh J.K.,
- Bailey K.R.,
- Nichols B.A.,
- Monahan K.H.,
- Tajik A.J.
- Maron B.J.,
- Gottdiener J.S.,
- Roberts W.C.,
- Henry W.L.,
- Savage D.D.,
- Epstein S.E.
- Maron B.J.,
- Harding A.M.,
- Spirito P.,
- Roberts W.C.,
- Waller B.F.
- Levine R.A.,
- Vlahakes G.J.,
- Lefebvre X.,
- et al.